> implies Pareto Efficient becomes a zero-sum game.
I’m not sure that is a meaningful (or indeed correct) conclusion. If, in the widget sale example we settle on a price of $2.45, we have a non-zero-sum transaction: I end up with $0.45 more value than my WTA, you end up paying $0.55 less than the widget is worth to you. Now, if the sales price is changed to $2.55, my gain of $0.10 is your loss of $0.10, but that makes the original transaction not zero sum.

I don’t think there is another optimum. There is a range of satisficing solutions, which may, or may not incorporate the ‘common good’ mental account. Now, of course, if that account is not there, it may be possible to “activate” it through communication to the agent. This would be akin to persuading a consumer that it is worth paying more for an otherwise functionally identical, but sustainable alternative.